Yahoo Fire Eagle

» see fireeagle.yahoo.net as it lived, on the wayback machine

~ the obituary ~

fire eagle never reached turkey at meaningful scale. it was a developer-platform product, in english, for an audience of u.s. early adopters who already knew what oauth was and already kept their phones publishing latitude. that overlap with the turkish web in 2008-2010 was thin. i was a graduate student then, on the engineer side of twitter, not the location-sharing side.

the project came out of yahoo brickhouse, a small san francisco team yahoo set up around 2007 to ship experimental things outside the main org. tom coates was the front-facing lead, with simon willison and mor naaman among the core engineers. the brief was to build a place where the user owns their location and decides which apps can read it.

the closed beta opened in late 2007. the public beta opened on march 4 2008. partner integrations spanned pownce and dopplr through brightkite and movable type. fire eagle was one of the first consumer-facing products to use oauth, which gives it a footnote in the history of permissioned api design. it was killed quietly in february 2013, almost three years after tom coates had already left yahoo in may 2010 to start his own product consultancy.

~ the case file ~

~ what it was ~

fire eagle was an api in front of a database row. the database row was your latitude and longitude, optionally with a friendlier place name. you wrote to the row through any app that knew about fire eagle. you read from it through any other app that knew about fire eagle, after granting that app permission through what fire eagle called authorization tokens and the rest of the industry would call oauth.

what made fire eagle interesting in 2008 was the decoupling. before fire eagle, every social app that wanted your location asked you for it directly and stored its own copy. fire eagle proposed a single canonical store with the user in control of who could read it. the model anticipated permissions infrastructure that ios and android would build into their operating systems a few years later.

~ the slide ~

the cracks were yahoo cracks. brickhouse was experimental but never integral. when yahoo started laying off and shedding peripheral projects in 2009-2010, brickhouse was an obvious early casualty. tom coates left in may 2010. the team scattered. the api kept running for almost three years on momentum with no roadmap, used by a smaller and smaller set of apps as they shut down or moved to their own location stacks. by 2013 yahoo turned it off without much ceremony.

~ the funny part ~

the funny part is that fire eagle was directionally right about everything. the user-owns-their-location model became the default for ios in 2011 and android shortly after, just baked into the operating system rather than offered as a third-party broker. oauth itself became the universal pattern for permissioned api access between consumer apps. fire eagle solved problems people were going to have, four years before they had them, in a wrapper that no one needed to use once apple solved the same problem for free at the os layer.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »